Flink + Kafka stream processing job

[kiyeonjeon21]

Priority: P1

Flink has compose files but no actual stream processing job.

What to do

• Create a Flink SQL job reading from clickstream-events Kafka topic
• Compute windowed aggregation (1-minute tumbling window: counts by action + device)
• Write results to new Kafka topic or ClickHouse via JDBC sink
• Document how to submit via Flink Dashboard at :8083

Acceptance criteria

• Flink job submits and runs without errors
• Aggregation results visible in the sink
• Flink Dashboard shows running job with checkpoints
• README explains submit/monitor/stop